计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (8): 35-39.

• 理论研究、研发设计 • 上一篇    下一篇

基于AHP-RBF的Swift云存储负载预测

谭  乾1,江  弋1,林  凡2   

  1. 1.厦门大学 信息科学与技术学院,福建 厦门 361005
    2.厦门大学 软件学院,福建 厦门 361005
  • 出版日期:2014-04-15 发布日期:2014-05-30

Load prediction of Swift cloud storage based on AHP-RBF

TAN Qian1, JIANG Yi1, LIN Fan2   

  1. 1.School of Information Science and Technology, Xiamen University, Xiamen, Fujian 361005, China
    2.School of Software, Xiamen University, Xiamen, Fujian 361005, China
  • Online:2014-04-15 Published:2014-05-30

摘要: 通过对Swift云存储中Proxy Node的负载因素研究,提出结合层次分析法(AHP)和混合递阶遗传训练的RBF神经网络实现对Swift云存储负载情况的预测,其中使用AHP构造对云存储系统的负载层次化模式,提高负载预测的综合精度,设计了RBF神经网络预测模型,用混合递阶遗传算法(HHGA)确定RBF神经网络的参数和结构。仿真实验结果表明,对Swift云存储负载的预测具有可行性,能为系统动态负载均衡决策提供依据。

关键词: Swift, 混合递阶遗传算法, 径向基函数(RBF)神经网络, 层次分析法, 负载

Abstract: Through the study of Proxy Node load factors in Swift cloud storage, a method which combines Analytic Hierarchy Process(AHP) and Hybrid Hierarchical Genetic Algorithm for training of Radial Basis Function Neural Network (HHGA-RBFNN) is proposed to predict Swift cloud storage load. This paper uses AHP to construct load hierarchy model of the system for raising comprehensive accuracy of load prediction of the system, designs RBFNN prediction model, and uses hybrid hierarchical genetic algorithm to train RBFNN’s parameters and configuration. From the experimental results, this method is effective, and can be a selection for Swift cloud system load balancing decision.

Key words: Swift, hybrid hierarchical genetic algorithm, Radial Basis Function(RBF) neural networks, Analytic Hierarchy Process(AHP), load